Detecting safety concerns via subscriber safety control (SSC) system

Abstract

Techniques are described that facilitate analyzing real-time communication sessions between a client device of a client account and a third-party device for the purpose of detecting safety concerns that may impact a client associated with the client device. The real-time communication sessions may include voice or text communications provisioned by a telecommunication service provider. In response to detecting a safety concern, notification data may be sent to a trusted device associated with the client account that alerts a trusted person of the inferred safety concern. In some examples, a safety concern may be inferred based on an irregular performance or non-performance of a regularly conducted real-time communication session. Moreover, communication session modification data may be sent to a client device involved in the real-time communication session for the purpose of terminating the real-time communication session based at least in part on safety concerns.

What is claimed: 
     
       1. One or more non-transitory computer-readable media storing computer-executable instructions, that when executed on one or more processors, cause the one or more processors to perform acts comprising:
 receiving, from a client device associated with a client of a client account, communication session data associated with a real-time communication session between the client device and a third-party device, the communication session data including at least session metadata associated with the real-time communication session; 
 retrieving, from a data store, a client data model associated with the client, the client data model including client profile data associated with the client; 
 analyzing the communication session data by identifying data patterns between the session metadata and the client data model; 
 generating a subscriber safety score for the client, based at least in part on correlation of the session metadata and the client data model; and 
 transmitting notification data to a trusted device of the client account based at least in part on the subscriber safety score being greater than or equal to a predetermined safety threshold. 
 
     
     
       2. The one or more non-transitory computer-readable media of  claim 1 , wherein the client profile data includes a list of permissible contacts and the session metadata includes a third-party device identifier of the third-party device, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 determining that the third-party device is omitted from the list of permissible contacts, based at least in part on the third-party device identifier, and 
 wherein, determining that the subscriber safety score is greater than or equal to the predetermined safety threshold is based at least in part on an omission of the third-party device on the list of permissible contacts. 
 
     
     
       3. The one or more non-transitory computer-readable media of  claim 1 , further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 retrieving, from the data store, one or more historical communication sessions conducted by the client device over a predetermined time interval, individual ones of the one or more historical communication sessions including historical session metadata that indicates a corresponding time of day, a day of a week, and a third-party device identifier; and 
 generating the client profile data based at least in part on the one or more historical communication sessions. 
 
     
     
       4. The one or more non-transitory computer-readable media of  claim 1 , further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 retrieving, from the data store, one or more historical communication sessions associated with the client over a predetermined time interval; and 
 generating at least one voice biometric template for the client based at least in part on the one or more historical communication sessions, and 
 wherein, the client profile data includes the at least one voice biometric template. 
 
     
     
       5. The one or more non-transitory computer-readable media of  claim 1 , further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 retrieving, from the data store, one or more historical communication sessions associated with the client over a predetermined time interval; 
 identifying at least one historical communication session of the one or more historical communication sessions that includes a historical interaction between the client device and third-party device; 
 capturing, from the one or more historical communication sessions, a voice biometric sample of the client from the at least one historical communication session; and 
 generating a voice biometric template of the client, based at least in part on the voice biometric sample, and 
 wherein, the client profile data includes the voice biometric template of the client. 
 
     
     
       6. The one or more non-transitory computer-readable media of  claim 1 , wherein the session metadata includes a third-party device identifier of the third-party device and a current geographic location of the client device, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 determining that a third-party associated with the third-party device is listed on a grey-list of contacts, based at least in part on the third-party device identifier, the grey-list of contacts corresponding to a list of contacts that the client is permitted limited communication based at least in part on one or more communication rules, the one or more communication rules including at least a particular geographic location; and 
 determining that the client is not permitted to interact with the third-party, based at least in part on the current geographic location of the client device, and 
 wherein, generating the subscriber safety score is further based at least in part on the client is not being permitted to interact with the third-party. 
 
     
     
       7. The one or more non-transitory computer-readable media of  claim 1 , wherein the communication session data further includes audio data, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 retrieving, from the data store, a list of impermissible vocabulary associated with the client; and 
 analyzing the audio data to detect a vocabulary used during the real-time communication session, and 
 wherein, the subscriber safety score is based at least in part on a comparison of the vocabulary used during the real-time communication session and the list of impermissible vocabulary. 
 
     
     
       8. The one or more non-transitory computer-readable media of  claim 1 , wherein the communication session data further includes audio data, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 analyzing the audio data to detect a current ambient noise that is associated with the client device during the real-time communication session, and 
 wherein, determining the subscriber safety score is further based at least in part on correlation of the current ambient noise with instances of historical ambient noise from the client profile data. 
 
     
     
       9. The one or more non-transitory computer-readable media of  claim 1 , wherein the session metadata further includes a current geographic location of the client device, and
 wherein, determining the subscriber safety score is further based at least in part on correlation of the current geographic location with historical geographic locations from the client profile data. 
 
     
     
       10. The one or more non-transitory computer-readable media of  claim 1 , further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 generating the notification data to include computer-executable instructions that automatically present an alert on a display of the trusted device that indicates an inferred safety concern for the client device. 
 
     
     
       11. The one or more non-transitory computer-readable media of  claim 10 , wherein the predetermined safety threshold is a first predetermined safety threshold, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 determining that the subscriber safety score is greater than a second predetermined safety threshold, the second predetermined safety threshold being greater than the first predetermined safety threshold; and 
 generating communication session modification data for delivery to the client device, the communication session modification data including computer-executable instructions that automatically terminate the real-time communication session. 
 
     
     
       12. The one or more non-transitory computer-readable media of  claim 1 , wherein the session metadata includes a geographic location of the client device at a point in time that the real-time communication session was initiated, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 retrieving, from the data store, one or more distribution rules associated with delivery of the notification data to the trusted device, the one or more distribution rules identifying a plurality of trusted devices based on the geographic location of the client device at the point in time that the real-time communication session was initiated; and 
 identifying the trusted device from the plurality of trusted devices for delivery of the notification data based at least in part on the one or more distribution rules and the session metadata. 
 
     
     
       13. A computer-implemented method, comprising:
 under control of one or more processors: 
 detecting that a client device is engaged in a real-time communication session with a third-party device, the client device being associated with a Secondary Account Holder (SAH) of a client account; 
 retrieving, from the client device, communication session data associated with the real-time communication session, the communication session data including audio data and session metadata of the real-time communication session, the session metadata including a current geographic location of the client device and a third-party device identifier associated with the third-party device; 
 retrieving, from a data store, a client data model associated with the SAH, the client data model correlating instances of historical communication sessions between the SAH and one or more third-parties; 
 analyzing the communication session data by identifying data patterns between the session metadata and data-points of the client data model; 
 inferring a subscriber safety score that is associated with the real-time communication session, based at least in part on correlation of the session metadata and the data-points of the client data model; and 
 transmitting notification data to a trusted device associated with the client account, based at least in part on the subscriber safety score. 
 
     
     
       14. The computer-implemented method of  claim 13 , further comprising:
 determining that the subscriber safety score is greater than a predetermined safety threshold; and 
 generating communication session modification data for delivery to the client device, the communication session modification data including computer-executable instructions that automatically terminate the real-time communication session, based at least in part the subscriber safety score being greater than the predetermined safety threshold, and 
 wherein the notification data includes computer-executable instructions that present an alert on a display of the trusted device, the alert indicating that a safety concern associated with the SAH has been inferred and the real-time communication session has been terminated. 
 
     
     
       15. The computer-implemented method of  claim 13 , further comprising:
 retrieving, from the data store, one or more historical communications sessions associated with the SAH over a predetermined time interval, the one or more historical communication sessions including corresponding session metadata and corresponding audio data; 
 generating a voice biometric template for the SAH based at least in part on the one or more historical communication sessions, the voice biometric template reflecting a baseline voice pattern for the SAH; and 
 generating the client data model to incorporate the voice biometric template. 
 
     
     
       16. The computer-implemented method of  claim 13 , further comprising:
 determining that the real-time communication session between the SAH and a third-party user associated with the third-party device is an irregular real-time communication session, based at least in part on correlation of the communication session data with data-points of the client data model; and 
 determining that the subscriber safety score is greater than or equal to a predetermined safety threshold, based at least in part on the real-time communication session being determined as the irregular real-time communication session. 
 
     
     
       17. A system comprising:
 one or more processors; 
 a memory coupled to the one or more processors, the memory including one or more modules that are executable by the one or more processors to: 
 receive, from a client device, communication session data associated with a real-time communication session between the client device and a third-party device, the communication session data including audio data and session metadata, and wherein the client device is associated with Secondary Account Holder (SAH) of a client account; 
 retrieve, from a data store, instances of historical communication sessions conducted by the SAH over a predetermined time interval, the instances of historical communication sessions including corresponding audio data and corresponding session metadata; 
 generate a client data model based on the instances of historical communication sessions; 
 correlate the audio data and the session metadata of the real-time communication session with data-points of the client data model; 
 determine a subscriber safety score based at least in part on a correlation of the audio data and the session metadata, and the data-points of the client data model; 
 determine whether the subscriber safety score is greater than or equal to a predetermined safety threshold; and 
 in response to the subscriber safety score being greater than the predetermined safety threshold, generate notification data for delivery to a trusted device, the notification data including computer-executable instructions that present an alert on a display of the trusted device that indicates a safety concern for the SAH has been inferred. 
 
     
     
       18. The system of  claim 17 , wherein the predetermined safety threshold is a first predetermined safety threshold, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 determine whether the subscriber safety score is greater than or equal to a second predetermined safety threshold, the second predetermined safety threshold being greater than the first predetermined safety threshold; and 
 in response to subscriber safety score being greater than the second predetermined safety threshold, generate communication session modification data for delivery to the client device, the communication session modification data including computer-executable instructions that automatically terminate the real-time communication session. 
 
     
     
       19. The system of  claim 17 , further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 determine whether the real-time communication session is an irregular real-time communication session, based at least in part on correlation of the session metadata and with the data-points of the client data model, and 
 wherein, to infer the subscriber safety score is based at least in part on a determination of whether the real-time communication session is the irregular real-time communication session. 
 
     
     
       20. The system of  claim 17 , wherein the session metadata includes a time of day and a day of week that the real-time communication session was initiated, and further storing instructions that, when executed cause the one or more processors to perform acts comprising:
 retrieve, from the data store, one or more distribution rules associated with delivery of the notification data to the trusted device, the one or more distribution rules identifying one or more trusted devices based on the time of day and the day of week that the real-time communication session was initiated; and 
 identify the trusted device from the one or more trusted devices for delivery of the notification data based at least in part on the one or more distribution rules and the session metadata.